Job Description & Details
This is a heavy-hitting hands-on AI role for someone who has moved past simple LLM wrapper demos and knows how to ship resilient, production-grade agentic systems on AWS. You will be knee-deep in architecting systems where failure handling, low latency, and high concurrency actually matter. If you enjoy building complex RAG pipelines and autonomous workflows that need to survive enterprise traffic, this one is worth a serious look.
What You'll Actually Be Doing
Expect to spend your days designing and scaling conversational interfaces and multi-agent workflows that talk to enterprise data sources without breaking a sweat. You will be writing clean Python code, configuring AWS Bedrock infrastructure, and wrangling vector databases to keep retrieval speeds blazing fast. Much of your time will go toward nailing down production deployment quirks—like implementing guardrails, setting up observability, and optimizing inference calls so they don't bottleneck under high load.
The Core Tech Stack
Python is your daily driver here, paired tightly with AWS Bedrock, Bedrock AgentCore, and core serverless infrastructure like Lambda, API Gateway, and ECS. You need deep, practical familiarity with vector databases like OpenSearch, Pinecone, or pgvector because semantic search performance will make or break your pipelines. If you haven't spent serious time orchestrating complex agent loops using tools like LangGraph, CrewAI, or building custom MCP servers, you'll find yourself struggling to keep up with the technical complexity of their architecture.
Interview Expectations
Be ready for the hiring panel to grill you on how you handle concurrency and latency bottlenecks when scaling LLM apps in production. They will likely ask you to debug a hypothetical failure scenario where an autonomous agent gets stuck in a recursive loop or times out against a downstream API, looking to see how you implement guardrails and fallback mechanisms. They aren't looking for textbook definitions; they want battle scars and practical mitigation strategies you've actually deployed.
Application Advice
Skip the generic AI buzzwords and tailor your resume to scream AWS infrastructure and production GenAI delivery. Make sure keywords like AWS Bedrock, LangGraph, vector databases, and specific cloud components (Lambda, ECS, DynamoDB) are front and center. Show them you know how to build secure, observable systems by highlighting any past experience with LLM evaluation frameworks, IAM policies, and cloud monitoring tools.